Relative efficiencies of some lr- estimators with type ii censored data and small samples

Document Type

Article

Publication Date

1-1-1994

Abstract

We consider several biased and unbiased LR estimators for the location and scale parameters of the normal, double exponential (or Laplace), logistic, and the smallest extreme value (SEV) distributions when the data are Type II censored. It is shown that Blom estimators, based on commonly recommended plotting positions, are highly efficient with respect to the best linear unbiased estimators (BLUE) when the assumed model is correct. But a commonly recommended Blom biased estimator has very low efficiencies in estimating the scale parameter of these distributions. We also show that some least squares estimators can be quite inefficient to estimate the parameters from non-normal distributions, and that the Winsorizcd and trimmed estimators are very robust when the data are contaminated by large outliers. © 1994, Taylor & Francis Group, LLC. All rights reserved.

Publication Source (Journal or Book title)

Journal of Statistical Computation and Simulation

First Page

151

Last Page

160

This document is currently not available here.

Share

COinS